一种高精确度的过程模型修复方法

发布时间：2018-02-02 07:30

本文关键词： 校准模型修复扩展校准修复算法　出处：《计算机集成制造系统》2017年05期 　论文类型：期刊论文

【摘要】：为了解决现有方法修复的过程模型精确度不高的问题,提出一种高精确度过程模型修复方法。为了便于确定偏差的位置,基于Petri网可达标识提出扩展校准的概念。针对扩展校准中的日志动作收集形成子日志,并将子日志挖掘出的子过程插入原模型中,避免了现有方法由于添加自环导致这一子过程多次重复发生的问题。结合Petri网的过程树,通过查找过程树的非叶子节点,能够定位到Petri网的选择结构。针对选择结构提出一种新的偏差类型,并给出判定方法,将挖掘出的子过程作为选择结构的一条分支,对模型进行修复。通过青岛某医院就诊数据的实例分析,验证了该修复算法有较高的拟合度和精确度。
[Abstract]:In order to solve the problem that the accuracy of the process model repaired by the existing methods is not high, a high precision process model repair method is proposed, in order to determine the position of deviation easily. The concept of extended calibration is proposed based on the reachability label of Petri net. The sub-log is collected and formed for the log action in extended calibration, and the sub-process mined by the sub-log is inserted into the original model. The existing method can avoid the problem that the sub-process is repeated many times because of adding self-ring. Combined with the process tree of Petri net, the non-leaf nodes of the process tree are searched. The selection structure of Petri net can be located. A new type of deviation is proposed for the selection structure, and the decision method is given. The subprocess mined out is regarded as a branch of the selection structure. The model was repaired by a case study of a hospital in Qingdao, and it was verified that the algorithm had high fitting degree and accuracy.
【作者单位】：山东科技大学计算机科学与工程学院;
【基金】：国家自然科学基金资助项目(61170078,61472228) 山东省泰山学者建设工程专项资助项目山东省自然科学基金资助项目(ZR2014FM009) 山东省优秀中青年科学家科研奖励基金资助项目(BS2015DX010) 山东科技大学科技创新资助项目(SDKDYC170224)~~
【分类号】：TP301.1
【正文快照】： 0引言过程挖掘是指那些从实际执行集合中提取出结构化过程描述的方法[1]。过程挖掘技术有过程发现、一致性检测和过程增强3类应用。其中,过程发现技术的目标是基于事件日志构建一个过程模型;一致性检测是将一个已知的过程模型与其产生的事件日志进行比较,以便检查记录在日志中

【相似文献】